Investor Protections And Concentrated Ownership:
Assessing Corporate Control Mechanisms
In The Netherlands
December 2002
Robert Chirinko, Hans van Ees,
Harry Garretsen, and Elmer Sterken
*
*
Emory and CESifo, Groningen, Utrecht and CESifo, and Groningen,
respectively. The authors acknowledge helpful conversations with and comments
from Kees Cools, Oscar Couwenberg, George Gelauff, Leo de Haan, Niels
Hermes, Rez Kabir, Huntley Schaller, and an anonymous referee, and seminar
participants at the American Economic Association (especially Wolfgang Eggert),
CESifo Area Conference on The Economics Of Organization And Corporate
Governance Structures (especially Josef Zechner), European Economic
Association, Groningen, the Institute For Advanced Studies (Vienna), Tilburg, and
Vienna. We thank Tutein Nolthenius Publishers for supplying data, and Emory's
Institute For Comparative And International Studies and University Research
Committee and Groningen's S.O.M. for partial financial support. All errors,
omissions, and conclusions remain the sole responsibility of the authors. This is a
revised version of CESifo Working Paper No. 210.
Investor Protections And Concentrated Ownership:
Assessing Corporate Control Mechanisms
In The Netherlands
Table Of Contents
Abstract
I.
Introduction
II.
The Corporate Control System In The Netherlands
A. Management And Supervisory Boards
B. Shareholders
C. Institutions
D. Summary
III.
The Dataset
IV.
Estimation Strategy
V.
The Substitution Hypothesis
VI.
Firm Performance, Investor Protections, And Concentrated Ownership
VII. Summary And Conclusions
References
Tables
Investor Protections And Concentrated Ownership:
Assessing Corporate Control Mechanisms
In The Netherlands
Abstract
R. Chirinko, H van Ees, H. Garretsen, and E. Sterken*
The Berle-Means problem -- information and incentive asymmetries
disrupting relations between knowledgeable managers and remote investors -- has
remained a durable issue engaging researchers since the 1930's. However, the
Berle-Means paradigm -- widely-dispersed, helpless investors facing strong,
entrenched managers -- is under stress in the wake of the cross-country evidence
presented by La Porta, Lopez-de-Silanes, Shleifer, and Vishny and their legal
approach to corporate control.
This paper continues to investigate the roles of investor protections and
concentrated ownership by examining firm behaviour in the Netherlands. Our
within country analysis generates two key results. First, the role of investor
protections emphasized in the legal approach is not sustained. Rather, we find that
performance is enhanced when the firm is freed of equity market constraints, a
result that we attribute to the relaxation of the myopia constraints imposed by
relatively uninformed investors. Second, ownership concentration does not have a
discernible impact on firm performance, which may reflect large shareholders'
dual role in lowering the costs of managerial agency problems but raising the
agency costs of expropriation.
JEL No.: G3
Professor Robert S. Chirinko
Department of Economics
Emory University
Atlanta, Georgia 30322-2240
U.S.A.
Professor Hans van Ees
Department of Economics
University of Groningen
P.O. Box 800
Groningen 9700 AV
THE NETHERLANDS
PH: 1-773-525-5410
FX: 1-773-525-0144
EM: rchirin@emory.edu
PH: 31-50-3633705
FX: 31-50-3637207
EM: h.van.ees@eco.rug.nl
Professor Harry Garretsen
Utrecht School of Economics
Utrecht University
Vredenburg
1383511 BG Utrecht
THE NETHERLANDS
Corresponding Author
Professor Elmer Sterken
Department of Economics
University of Groningen
P.O. Box 800
Groningen 9700 AV
THE NETHERLANDS
*
PH: 31-30-253 9810
EM: h.garretsen@econ.uu.nl
PH: 31-50-3633723
FX: 31-50-3637207
EM: e.sterken@eco.rug.nl
Investor Protections And Concentrated Ownership:
Assessing Corporate Control Mechanisms
In The Netherlands
But have we any justification for assuming that those in
control of a modern corporation will also choose to
operate it in the interests of the owners?
Berle and Means (1932, p. 113)
...the stockholders' position, once a controlling factor in
the running of the enterprise, has declined from extreme
strength to practical impotence.
Berle and Means (1932, p. 131)
I. Introduction
The Berle-Means problem -- information and incentive asymmetries
disrupting relations between knowledgeable managers and remote investors -- has
remained a durable issue engaging researchers since the 1930's. However, the
Berle-Means paradigm -- widely-dispersed, helpless investors facing strong,
entrenched managers -- is under stress and of only modest applicability for most
developed countries. In a recent paper, La Porta, Lopez-de-Silanes, and Shleifer
(LLS, 1999), examine the 20 largest firms in 27 wealthy countries, and show that
concentrated ownership is the norm, not the exception.1 The Berle-Means
paradigm is appropriate for the United States and United Kingdom, countries
whose legal environments provide very good protection of minority shareholders'
rights. LLS' cross-country results emphasize that investor protections and
concentrated ownership are the key elements for understanding how the modern
1
See Barca and Becht (2001) and Gugler (2001) for further analysis of concentrated ownership.
2
corporation is controlled and "how to assure financiers that they get a return on
their financial investment" (Shleifer and Vishny, 1997, p. 773).2
This paper continues to investigate these two mechanisms of corporate
control by examining the roles of investor protections and concentrated ownership
on firm behaviour in the Netherlands. A within-country analysis complements the
prior cross-country research because several factors -- taxes and regulations -- can
be held constant. Dutch firms are very useful for studying these mechanisms of
control because several devices exist for circumscribing investor protections that
are used by many, but not all, firms. Ownership concentration also ranges widely.
The goal of this paper is to assess the impact of corporate control
mechanisms on firm performance. Section II discusses the system of corporate
control in the Netherlands, and reviews the roles of three key sets of actors:
management and supervisory boards, shareowners, and institutions. The financial
statement, anti-investor protection, and ownership data are available for 93 firms
for the period 1992 to 1996. Section III discusses these data, and documents the
variation in anti-investor protections and concentrated ownership in the sample.
Section IV presents the estimation strategy for assessing the impact of antiinvestor protections and concentrated ownership as corporate control mechanisms.
We discuss the ambiguity of these control mechanisms on long-run performance
(measured by the long-run return on assets computed as a five-year average), and
the appropriateness of focusing on the cross-section dimension of the data.
Empirical results are contained in the next two sections. Section V contains
a simple test of the substitution hypothesis between anti-investor protections and
ownership concentration that is an important element in the legal approach to
2
Roe (1994) presents an alternative and provocative theory of how political forces created a
dispersed structure of corporate ownership in the United States. See Carney (1997) and LLS (1998,
Section III) for critiques, and Roe (1999) for further elaboration of his thesis.
3
corporate governance. We find no support for the substitution hypothesis in our
Dutch data. Rather, it seems that once anti-investor protections are in place, the
firm is either avoided or abandoned by concentrated owners. Regression results
are contained in Section VI. We uncover a strong positive relation between the
absence of investor protections and profitability. That is, as firms are freed from
the pressure of equity markets, profitability rises. No systematic relation is
uncovered for ownership concentration.
A summary and conclusions are presented in Section VII. In interpreting
their results presented in this paper, it is important to keep in mind that this
evidence is more in the spirit of interesting conditional correlations than definitive
tests about fundamental causes. The current research aims to stimulate and direct
theoretical modeling that will ultimately provide the basis for more formal
hypothesis testing. Thus, our empirical evidence is intended to inform views of
corporate control problems and the solutions provided by various control
mechanisms.
4
II. The Corporate Control System In The Netherlands
The Dutch system of corporate control contains three key sets of actors:
3
management and supervisory boards, shareowners, and institutions.
A. Management And Supervisory Boards
The focal point of the corporate control system is a two-tier board structure
consisting of a management board (Raad van Bestuur) in charge of the day-to-day
operations of the firm and a supervisory board (Raad van Commissarissen). The
supervisory board's scope of influence varies substantially depending on which
organizational regime the firm adopts. The structural regime (Structuurregeling)
described here applies to the majority of public limited liability companies
(Naamloze Vennootschappen, NV's) listed on the Amsterdam Stock Exchange.
The supervisory board has three primary functions: to appoint (usually for an
indefinite term), monitor, and dismiss members of the management board (though
4
the latter rarely occurs); draft the annual financial statement for presentation at
the annual shareholders meeting; and approve major business decisions proposed
by the management board concerning, for example, expansions, acquisitions,
restructurings, or financing.
Members of the supervisory board are appointed for four year terms by co5
option, that is, by the incumbent members of the supervisory board. An
3
Detailed descriptions of the system of corporate control and finance in the Netherlands can be
found in CPB (1997, Chapter 10), Gelauff and den Broeder (1997), and de Jong, Kabir, Marra, and
Röell (2001). We have relied particularly heavily on Gelauff and den Broeder's work in writing this
section.
4
Dutch management board turnover, calculated as the number of management board members
leaving the firm by other than natural causes and scaled by board size, is approximately 8.0% (van
Oijen, 2000). This figure is somewhat low compared to those reported for other countries; Kaplan
(1994) reports turnover rates of 12.0% (excluding cases of death and illness) for the United States
and 10.0% for Germany.
5
The mean [median] number of members on the supervisory and management boards for our
5
individual can not serve on both the supervisory and management boards of the
same company. In practice, the management board has a very large influence on
appointments to the supervisory board (van der Goot and van het Kaar, 1997).6
The two-tier board structure in the Netherlands differs substantially from that in
Germany, where the supervisory board is appointed by the shareholders at the
annual meeting and exerts substantial independent influence on management. The
close relations between management and supervisory boards makes the Dutch
two-tier system somewhat similar to the U.S. system, where executive managers
sit on the board of directors (comparable to the supervisory board) and the chief
executive officer often chairs the board of directors. In sum, the Dutch
supervisory board is largely advisory, though that counsel may receive more
attention depending on the background of the advising member.
B. Shareholders
Shareholders/investors exercise control through voting at the annual
meeting (Algemene Vergadering van Aandeelhouders) and, for large investors, by
sitting on the supervisory board. At the annual meeting, the influence of investors
is circumscribed under the structural regime in two ways. First, few important
issues come before the annual meeting: the financial statement drafted by the
supervisory board is voted on (amendments are not permitted), nominations for the
supervisory board may be proposed and rejected, though election is by incumbent
members of the supervisory board. Large investors can exert influence by refusing
sample of firms is 4.95 [5] and 2.95 [3], respectively.
6
Inside (managerial) ownership of listed firms is unimportant in the Netherlands. Based on an
ownership criterion of 5.0%, 19 of 137 firms were owned by insiders sitting on the management
board and an additional and different 6 firms for those on the supervisory board (de Jong et. al.
(2001, Table 11). Majority ownership was achieved by the management board of 5 firms and by
the supervisory board of 1 firm. For the 137 firms, average ownership is 4.86% and 1.82% for
management and supervisory boards, respectively.
6
to approve the financial statements and supervisory board nominations.
Second and more devastating to investors' voting rights, management has
available four potent devices for diluting voting power and separating control
rights from cash flow rights:
--
Preference shares ('prefs') have the same voting rights as ordinary
equity, but have a fixed payout with priority over payouts of ordinary
dividends. These shares are frequently held by a continuity
foundation (Stichting Continuiteit) friendly to the firm's management.
Preference shares, which can be issued on a temporary basis,
effectively increase the voting power of managers. This anti-investor
protection (AIP(1)) is used by 66.0% of the firms in our sample.
--
Tradable depository receipts (TDR's) also separate cash flow and
control rights. Under this procedure, the ordinary equity capital is
deposited at an administrative office (Administratie-kantoor), and
TDR's are issued (similar to American Depository Receipts issued on
non-U.S. equity). TDR's generally entitle the holder to cash flow
rights, but control rights reside with the administrative office. This
anti-investor protection (AIP(2)) is used by 32.0% of the firms in our
sample.
--
Priority shares (Prioriteits-aandelen) carry special voting rights on
matters such as "proposing or preventing the appointment of
particular new members of the management and supervisory boards,
approving the issue of ordinary shares, liquidation of the company or
changing the articles of association" (Gelauff and den Broeder, 1997,
p. 67). The issuance of priority shares curtails voting power for
extant shareholders. This anti-investor protection (AIP(3)) is used by
24.0% of the firms in our sample.
In the face of these three AIP's, shareholders have little reason to pursue
aggressively their limited tasks granted under the structural regime. The Dutch
annual meeting differs radically from its German and U.S. counterparts, where, in
7
principle, shareholders have a powerful effect on the course of events primarily by
electing the supervisory board (or board of directors) and voting on important
matters brought before shareholders.
The above AIP's all focus on altering effective voting rights by issuing
shares. In considering voting rights in the Netherlands, it is important to bear in
mind that investor protections and the tasks voted upon at the annual meeting are
directly linked to the applicable organizational regime. Under other organizational
regimes available to Dutch firms, shareholders exercise much more influence. A
fourth AIP is created by the option that allows firms to choose to be governed
according to the rules and regulations of the structural regime described above
7
even if they are not required to do so by law.
Investor protections are enhanced under the other two other legal regimes
relevant to public limited liability companies. Firms that meet the criteria for the
structural regime but are majority foreign-owned can follow the mitigated
structural regime (Gewijzigde Structuurregeling). Under this legal regime, the
supervisory board's responsibilities for appointing and dismissing members of the
management board and drafting the annual financial statement are transferred to
the annual shareholders meeting, enhancing investor protections. Public limited
liability companies that do not meet the above criteria can adopt the common legal
regime (Vennootschapsrecht) for which a supervisory board is optional. If a
supervisory board is in place, appointments are determined at the annual meeting,
8
and its only major responsibility is to approve major management decisions. All
7
The majority of Dutch firms listed on the Amsterdam Stock Exchange are required to follow the
structural regime because they satisfy all of the following conditions: the firm is a public limited
liability company, subscribed capital exceeds 25 million guilders (approximately $12.5 million
U.S.), employment in the Netherlands exceeds 100, and employees are represented by a works
council.
8
A survey of 180 smaller limited liability companies (with 50-1000 employees) -- who were not
8
other important decisions, especially the appointment of the management board,
are made at the annual shareholders meeting. In sum, investor protections are
substantially enhanced under the mitigated structural and common regimes, which
transfer power from the supervisory board to the annual meeting. The voluntary
adoption of the structural regime can thus be viewed as another device weakening
investor protections because the structural regime's supervisory board is largely
influenced by and hence empowers management at the expense of investors. This
anti-investor protection (AIP(4)) is used by 25.0% of the firms in our sample.
C. Institutions
The third set of key actors in Dutch corporate governance is Institutions -financial intermediaries, workers councils, and informal networks. Financial
Intermediaries also hold equity positions and, as discussed above, shareownership
per se may have little impact on controlling managers. However, their equity
stakes are occasionally large, and they are considered long-term, "patient"
investors. Consequently, financial intermediaries frequently obtain seats on the
9
supervisory board. Furthermore, banks are actively involved in extending shortterm credit, and thus have a direct and potentially powerful channel of influence
on management (cf. Shleifer and Vishny, 1997, Section IV.C). The role of Dutch
banks is much greater than in the United States, where banks are largely prohibited
from owning equity and, until very recently, were small by the standards of
Continental Europe. By contrast, banks have a long-standing and prominent role
legally obligated to install a supervisory board but did so voluntarily -- revealed that the main
motive for the installation was the need for expert advice (Gelauff and den Broeder, 1997, pp. 4445).
9
Pension funds and insurance companies own more equity in the Netherlands than in Germany
(13.4% vs. 7.1%) but much less than their U.S. counterparts (24.7%). (Data are for 1993, and are
taken from Gelauff and den Broeder, 1997, p. 46).
9
on the corporate landscape in Germany where they hold large positions in both
debt and equity and actively serve on, and frequently chair, supervisory boards.
Employees are represented by a Workers Council (Ondernemings-raad) that
is voluntary but exists at virtually all large firms. The works council has some
influence with and occasional membership on the supervisory board, where
usually one member represents workers' concerns. The works council has the
same rights as shareholders to propose or reject nominations to the supervisory
board. The position of the works councils bear some resemblance to that played
by organized labour in the United States, where union representatives frequently
hold a seat on the board of directors. By contrast, legal statutes grant German
workers much more nominal influence on corporate affairs, including between
one-third to one-half of the seats on the supervisory board.
10
However, the chair
of the German supervisory board holds the tie-breaking vote, and this position is
usually held by a person (frequently a banker) sympathetic to management's
concerns.
Networks of outside board members are potentially importantly for control.
These individuals hold positions on the management and supervisory boards of
several companies, and/or they are "distinguished experts" drawn from the ranks
of politicians, civil servants, lawyers, professors, and former directors. With their
perspective and experience, these outside board members may provide valuable
advice to firms.
D. Summary
Dutch firms are affected by a variety of different corporate control
mechanisms. The extent of investor protections and concentrated ownership vary
10
German co-determination laws require that, for stock companies with 500 or more employees,
one-third of the seats on the supervisory board must be held by persons elected by the employees.
The fraction increases to one-half for stock companies with 2,000 or more employees.
10
across firms, and this variation will be used in the empirical work to assess the
impacts of these corporate control mechanisms on firm performance. Financial
intermediaries, workers councils, and informal networks are additional control
mechanisms that may also affect corporate governance and firm performance, but
they will not be analyzed further in this study.
11
III. The Dataset
We draw on two sources to construct the variables used in our empirical
analysis. Our sample extends from 1992-1996, unless otherwise noted. Financial
statement data are obtained from the AMADEUS/REACH database covering 165
Dutch firms. We focus only on firms involved in manufacturing, omitting
financial and service firms. The data will be time-averaged to form the crosssection used in the econometric work. Thus, a balanced panel is highly desirable,
and we omit firms involved in mergers or takeovers. Two firms are excluded:
Hunter Douglas because of erratic data and Royal Dutch Shell because it is
registered in both the United Kingdom and the Netherlands. These restrictions
lead to a final sample consisting of 93 firms. Variables based on financial
statement data are as follows:
CFA
=
Cash flow (operating income plus depreciation) / TA;
CVCFA
=
Coefficient of variation of CFA (the standard deviation
divided by the mean);
DIVERSITY
=
The percentage of two-digit industrial activity in which
the firm is involved outside its core business, measured
by revenues;
LEVERAGE
=
(TA less stockholders equity) / TA;
PROF
=
Profitability (before-tax profits plus financial expenses)
/ TA;
SIZE
=
The natural logarithm of TA;
TA
=
Total assets less depreciation.
12
For each of the 93 firms, these variables are averaged from 1992 to 1996.
11
The
medians (m), means (µ), and standard deviations (σ) are presented in Table I.
.
Data for anti-investor protections (AIP's) and ownership concentration
(OWN's) for the firm are obtained from the Monitoring Commissie Corporate
Governance (1998). The AIP's and two of the ownership variables enter as
indicator variables. In those rare cases where the defining characteristic (e.g.,
issuing preference shares) changes over the sample, the indicator is determined by
the most frequent value of the characteristic. The ownership stake (OWN(L)) is a
continuous variable. The variables are defined as follows:
AIP(1)
=
1 if a firm issues preference shares;
0 elsewise.
AIP(2)
=
1 if the percentage of shares issued as tradable
depository receipts is equal to or greater than 50.0%; 0
elsewise.
AIP(3)
=
1 if a firm issues priority shares;
0 elsewise.
AIP(4)
=
1 if a firm is not required to implement the structural
regime, but does so voluntarily;
0 elsewise.
OWN(L)
=
Ownership stake of the largest shareholder stated as a
percentage of all outstanding ordinary equity.
OWN(20)
=
1 if OWN(L) > 0.20; 0 elsewise.
OWN(40)
=
1 if OWN(L) > 0.40; 0 elsewise.
11
For those series analyzed as ratios, the ratios are computed and then summed over time; that is,
ratioi = Σt (ai,t / bi,t). All of the ratios are defined so that bi,t is far from zero. This procedure
obviates the need for price deflators, which are unavailable on a firm-specific basis.
13
As shown in the first row of Table II, the largest shareholder (as measured
by (OWN(L)) controls more than 20% of outstanding ordinary equity for over onehalf of the firms. Fifteen percent of the firms are majority owned. Table II also
includes information on the distribution of the AIP's across categories of
ownership concentration. The first row contains the distribution for all 93 firms,
Relative to this benchmark, the different AIP's are used more or less in similar
proportions.
14
IV. Estimation Strategy
Our estimation strategy is to assess corporate control mechanisms by their
impact on economic performance. The links between good governance and good
performance are well established. As argued by Berle and Means, the incentives
of managers controlling the firm differ from those of owners and, in large, publicly
held corporations, most owners/investors have little incentive to expend resources
to ensure that the firm is operated in their interests. Consequently, resources may
be directed to activities that benefit managers rather then firm performance.
Jensen (1986) has emphasized that these agency problems are likely to become
particularly severe when managers have an abundance of internal resources
relative to investment opportunities (i.e., "Free Cash Flow"). The important
consequence is that the Berle-Means agency problem may lead the firm to a suboptimal allocation of resources that compromise performance.
Investment decisions within the firm may also be adversely affected by
undue pressure from equity investors. The "Managerial Myopia" model argues
that equity markets may not allocate capital efficiently because of an absence of
stable, dedicated investors with a long-term interest in the firm's performance.
Formal models of strategic behaviour establish that, when inside managers know
more about the firm's operations than outside owners, high hurdle rates that distort
investment can occur because of a premium for signal jamming, obfuscation, or
hidden action. Rather than being ameliorative, equity market pressures force
management into an undue focus on boosting short-term earnings and avoiding
takeover threats at the expense of long-run performance.12
In sum, agency and myopia problems adversely affect firms, and corporate
12
For discussions of capital allocation systems suffering from short-termism, see de Jong (1996)
for the Netherlands and Porter (1992) for the United States. See Bohlin (1997) for an excellent
survey of formal models of managerial incentives and investment biases.
15
control mechanisms that reduce the impact of these problems should be associated
with better performance. However, as the myopia model suggests, the link
between investor protections and performance is ambiguous. While a more active
role for investors may mitigate agency conflicts, myopia problems may be
exacerbated.
A similar ambiguity affects concentrated ownership. Concentrated
ownership is a direct solution to the agency problems that arise from the
separation of ownership from control. With a sufficiently large equity stake in the
firm, investors have the incentive to invest resources in monitoring and
disciplining managers and thus reducing agency problems. However,
concentrated owners may also use their controlling position to expropriate benefits
for themselves at the expense of small shareholders by, for example, diverting
resources towards other firms in which they have substantial cash flow rights.
Thus, the lower costs of managerial agency problems will be balanced against the
higher agency costs of expropriation. As a result of these conflicting tendencies,
the effects of AIP's and concentrated ownership on firm performance are
ambiguous theoretically. Their ultimate impact is an empirical matter.
Firm performance is measured by the long-run return on assets computed as
a five-year average. This computation exhausts the time dimension of our data,
and thus our empirical work is based on cross-section analyses. Since the
questions motivating this study focus on long-run relations among profitability
and various corporate control mechanisms, focusing on the cross-section
dimension of the data is appropriate. Moreover, there is very little time-series
variation in the AIP's and concentrated ownership. As a consequence of the nearconstancy of the corporate control variables over our sample period, we treat them
16
as exogenous regressors.13
Finally, financial market data do not provide a good measure of the impact of
corporate control variables on economic performance. In a sense, asset prices are
too good a measure because they capitalize the effects of favorable/unfavorable
corporate governance policies. Hence, the financial return for a firm serving
shareholders' interests may differ little from the financial return for a firm with
serious and unresolved corporate governance problems. The Brainard-Tobin Q is
also not a useful measure because it gravitates toward its equilibrium value of
unity, and it will be difficult to discern the impacts of control variables. Financial
data would be useful to examine firms switching control mechanisms over time
but, such time series variation and the associated event studies are not available to
us in this dataset.
13
A deeper understanding of corporate control issues requires us to model the adoption of
corporate control mechanisms, but this important task is not tackled in this paper.
17
V. The Substitution Hypothesis
An important implication of the legal approach to corporate governance is
that concentrated ownership substitutes for poor investor protections. This
relation has been confirmed in the cross-country comparisons reported in LLSV
(1998) and LLS (1999). For our within-country analysis, the comparable
implication is that
the AIP's will be positively correlated with ownership concentration.
Table III evaluates this hypotheses in terms of correlation coefficients for
various AIP's and three measures of ownership concentration -- the value of the
largest stake (OWN(L)) and indicator variables set to unity if the largest stake
exceeds 20.0% (OWN(20)) or 40.0% (OWN(40)). There is scant evidence of a
positive relation.14 Rather, in the case of preference shares or tradable depository
receipts, the correlations are negative and significant at the 1.0% level. Thus, the
substitution hypothesis receives little support.
The above statement of the substitution hypothesis may not be sufficiently
sensitive to a key maintained assumption. Who moves first: concentrated owners
or management boards?15 If concentrated owners move first by taking a dominant
equity position, the anticipated correlation in the data would be reversed. Given
that concentrated owners have an incentive to adopt good corporate governance
practices, we might expect the AIP's to be eliminated. Under this scenario,
14
The reported results for AIP(2) may be biased toward a positive correlation. Under AIP(2), the
percentage of shares held by an administrative office may be sufficiently large that the firm is
also classified as a concentrated owner. In this case, AIP(2) and OWN(L) would be positively
correlated, though such a correlation does not reflect a substitution of governance mechanisms.
This bias does not affect the conclusion to be drawn about the substitution hypothesis in Table
III.
15
We thank our reviewer for highlighting the importance of these additional assumptions. .
18
ownership concentration would be negatively correlated with the AIP's.
As indicated by the analysis of the Dutch corporate governance system in
Section II, management boards have a great deal of autonomy in adopting antiinvestor protections, and it appears more accurate to assume that management
boards "move first." Under this assumption, two scenarios might then unfold.
First, consistent with the substitution hypothesis, a firm shielded from investors by
the AIP's might be very inefficient, and hence tempting for a owner to accumulate
a large equity stake so that the potential efficiency gains are realized. This
scenario would yield a positive correlation between AIP's and concentrated
owners. Second, outside investors, regardless of the size of their equity stake, may
find AIP-shielded firm undesirable. Under this scenario, potentially large
shareholders would avoid and existing large shareholders would abandon the firm,
and we would expect a negative correlation between the AIP's and concentrated
ownership.
While simple correlations presented in Table III should not be
overinterpreted, it appears that once managers compromise investor protections
with AIP's , large shareholders either avoid or abandon the firm to its managers.
In this initial analysis, concentrated ownership does not appear to play an
ameliorative role in Dutch corporate governance.
19
VI. Firm Performance, Investor Protections, And Concentrated Ownership
The roles of investor protections and concentrated ownership as corporate
governance mechanisms are assessed in terms of their impact on long-run
profitability. We begin our analysis by establishing a benchmark regression
without any governing mechanisms but containing several conditioning variables - the coefficient of variation of cash flow (CVCFA), a measure of the diversity of
the firm's lines of business (DIVERSITY), leverage (LEVERAGE), a measure of
firm size (SIZE), and seven industry dummies (not reported).16 All variables are
five-year averages. As shown in column 1 of Table IV, only LEVERAGE is
statistically significant at conventional levels (as well as most of the industry
dummies), an effect that can be interpreted as reflecting the adverse effects of
finance constraints in external capital markets or as a signal of maturity (since
older firms tend to be both more highly levered and less profitable than younger
firms).
The remaining entries in Table IV examine the impact of three anti-investor
protections that depend on share issuance -- preference shares (AIP(1)), tradable
depository receipts (AIP(2)), or priority shares (AIP(3)). We begin by allowing
for the broadest possible scope for the AIP's; if a firm uses any one of these three
AIP's, then a dummy variable is one. As shown in column 2, the AIP's have a
substantial positive impact on firm performance that is both statistically and
economically significant. In the latter regard, firms that have one or more of these
AIP's are 25% more profitable than the complimentary class of firms.17 The
remaining entries consider two of the AIP's at time, and the issuance of priority
shares (AIP(3)) has the most consistent positive impact on profitability.
16
The explicit equation specification is presented in the notes to Tables IV-VI.
17
One-third of the firms in our sample do not use AIP(1), AIP(2), or AIP(3).
20
Table V evaluates the impact of the fourth AIP, the voluntarily adoption of
the structural regime (AIP(4)). As shown in column 1, firms that transfer power
away from the annual meeting to the management and supervisory boards are onethird more profitable than the average sample firm. Similar statistically significant
results hold in columns 2-4, where AIP(4) is interacted with the other AIP's.
The precisely estimated and economically important positive effects of
AIP's on profitability run counter to the legal approach to corporate governance.
The results in Tables IV and V cast some doubt on the prevailing view that robust
stock markets are the key to robust firm performance.
Many papers have argued that ownership concentration will attenuate
agency problems. Given the wide variation in ownership concentration in the
Netherlands (cf. Table II), this issue can be explored with our data. Table VI
contains several specifications using different measures of ownership
concentration -- the ownership stake of the largest shareholder stated as a
percentage of all outstanding ordinary equity (OWN(L)) entered as a linear term,
as both linear and quadratic terms or as indicator variables, OWN(20), or
OWN(40). For the Dutch firms in our sample, there is little systematic relation
between ownership concentration and firm performance, and little support for the
proposition that ownership concentration attenuates agency problems.
21
V. Summary And Conclusions
Recent research in corporate governance has reaffirmed the importance of
the Berle-Means problem of pervasive agency problems, but has also exposed an
important fault line in the Berle-Means paradigm -- the assumption of small and
uninformed investors. The emerging legal approach to corporate governance
highlights that the assumption of small investors is not applicable in most
industrialized countries, and emphasizes the importance of investor protections
and ownership concentration as complementary means for solving governance
problems. To date, most of the empirical work has been cross-country.
This paper uses the interesting institutional features in the Dutch economy
to undertake a within country analysis. Two key conclusions emerge. First, the
role of investor protections emphasized in the legal approach is not sustained.
Rather, we find that performance is enhanced when the firm is freed of equity
market constraints. We attribute this to the relaxation of the myopia constraints
that can be imposed by relatively uninformed investors. Second, ownership
concentration does not have a discernible impact on firm performance. Our results
indicate that concentrated ownership substituted with AIP's, a result consistent
with potentially large shareholders avoiding or abandoning firms with antiinvestor protections. In regression models of profitability, we find no systematic
effect of concentrated ownership, which may reflect the tension between the lower
costs of managerial agency problems versus the higher agency costs of
expropriation.
These conclusions concerning investor protections and concentrated
ownership should be viewed as preliminary. Two important caveats remain. First,
as is well know, results for one country may not be applicable outside its borders.
Second, we need a better understanding of the forces leading to the adoption of
22
AIP's and concentrated ownership. In this study, the adoption was taken as
exogenous because of the stability of the corporate control variables over our
sample. Nonetheless, we would like to have a better appreciation of the forces at
work influencing adoption decisions.
23
REFERENCES
Barca, Fabrizio, and Becht, Mario (eds.), The Control Of Corporate Europe
(Oxford: Oxford University Press, 2001).
Berle, Adolf A., Jr., and Means, Gardiner C., The Modern Corporation and
Private Property (New York: MacMillan, 1932). Revised edition published by
Harcourt, Brace & World in 1968.
Bohlin, Erik, "A Survey of Managerial Incentives and Investment Bias -Common Structure but Differing Assumptions," Journal Of Business Finance &
Accounting 24 (March 1997), 197-248.
Carney, William J., "Large Bank Stockholders In Germany: Saviors Or
Substitutes?," Journal of Applied Corporate Finance 9 (Winter 1997), 74-81.
CPB, Netherlands Bureau For Economic Policy Analysis, Challenging
Neighbors (Heildelberg: Springer, 1997).
Gelauff, G.M.M. and den Broeder, C., Governance of Stakeholder
Relationships: The German and Dutch Experience (Société Universitaire
Européenne de Recherches Financières Studies No. 1 (January 1997).
Goot, L.R.T. van der, and Kaar, R.H. van het, "Beschermingsconstructies,
Structuurregime En Ondernemings-waarde," Maandblad voor Accountancy en
Bedrijfseconomie 71 (1997), 497-507.
Gugler, Klaus, Corporate Governance and Economic Performance,
(Oxford: Oxford University Press, 2001).
Het Financieele Dagblad, Wet Melding Zeggenschapsrecht (WMZ) (various
issues).
Hofstede, Geert, Culture's Consequences (New York: Sage, 1980).
Inglehart, Ronald, Basañez, Miguel, and Moreno, Alejandro, Human Values
and Beliefs: A Cross-Cultural Sourcebook (Ann Arbor: University of Michigan
Press, 1998).
24
Jensen, Michael C., "Agency Costs of Free Cash Flow, Corporate Finance,
and Takeovers," American Economic Review 76 (May 1986), 323-329.
Jong, Abe de, Kabir, Rezaul, Marra, Teye, and Röell, Ailsa, "Ownership and
Control in the Netherlands," in Fabrizio Barca and Mario Becht (eds.), The
Control Of Corporate Europe (Oxford: Oxford University Press, 2001), 188-206.
Jong, H.W. de, "Rijnlandse Ondernemingen Presteren Beter," EconomischStatistische Berichten 13 (March 1996), 228-232.
Kaplan, Steven N., "Top Executive Rewards And Firm Performance: A
Comparison Of Japan And The United States," Journal of Political Economy 102
(1994), 510-546.
La Porta, Rafael, Lopez-de-Silanes, Florencio, and Shleifer, Andrei,
"Corporate Ownership Around the World," Journal of Finance 54 (April 1999),
471-518.
La Porta, Rafael, Lopez-de-Silanes, Florencio, Shleifer, Andrei, and Vishny,
Robert W., "Trust In Large Organizations," American Economic Review 87 (May
1997), 333-338.
La Porta, Rafael, Lopez-de-Silanes, Florencio, Shleifer, Andrei, and Vishny,
Robert W., "Law and Finance," Journal of Political Economy 106 (December
1998), 1113-1155.
Monitoring Commissie Corporate Governance, Monitoring Corporate
Governance in Nederland (Deventer: Kluwer, 1998).
Oijen, P. van, "Corporate Governance in the Netherlands," Unpublished
Ph.D. Thesis, University of Amsterdam (2000).
Porter, Michael E., Capital Choices: Changing The Way America Invests In
Industry (Washington: Council on Competitiveness, 1992).
Roe, Mark J., Strong Managers, Weak Owners: The Political Roots of
American Corporate Finance (Princeton: Princeton University Press, 1994).
25
Roe, Mark J., "Political Preconditions To Separating Ownership From
Control: The Incompatibility Of The American Public Firm With Social
Democracy," Columbia University School of Law (1999).
Shleifer, Andrei, and Vishny, Robert W., "A Survey of Corporate
Governance," Journal of Finance 52 (June 1997), 737-783.
26
TABLE I
SUMMARY STATISTICS
The entries are the median (m), mean (µ), and standard deviation (σ) of the indicated variable
averaged from 1992 to 1996 for each of the 93 firms. CFA is cash flow (operating income plus
depreciation) / TA, where TA is total assets less depreciation; CVCFA is the coefficient of
variation (the standard deviation divided by the mean) of CFA; DIVERSITY is the percentage of
two-digit industrial activity in which the firm is involved outside its core business, measured by
revenues; LEV is (TA less stockholders equity) / TA; OWN(L) is the ownership stake of the
largest shareholder stated as a percentage of all outstanding ordinary equity; PROF is profitability
(before-tax profits plus financial expenses) / TA; SIZE is the natural logarithm of TA.
Variable
CFA
CFACV
DIVERSITY
LEV
OWN(L)
PROF
SIZE
Median (m)
(1)
75.636
0.216
1.700
0.620
15.000
9.072
5.972
Mean (µ)
(2)
406.331
0.286
1.988
0.605
25.118
9.727
6.094
Standard Deviation (σ)
(3)
1214.386
0.286
2.052
0.124
21.920
4.775
1.655
27
TABLE II
DISTRIBUTION OF ANTI-INVESTOR PROTECTIONS
ACROSS LEVELS OF OWNERSHIP CONCENTRATION
The entries are the number of firms that use a given anti-investor protection (AIP). Column 1
contains the total number of firms using a given AIP; the remaining entries in a given row
distribute this total across levels of ownership concentration. All of the AIP's are indicator
variables taking a value of 1 for the criteria below and zero elsewise: AIP(1)=1 if a firm issues
preference shares; AIP(2)=1 if the percentage of shares issued as tradable depository receipts is
equal to or greater than 50%; AIP(3)=1 if a firm issues priority shares; AIP(4)=1 if a firm is not
required to implement the structural regime, but does so voluntarily. See Section II.B for further
discussion of the AIP's. Ownership concentration is measured by the ownership stake of the
largest shareholder stated as a percentage of all outstanding ordinary equity (OWN(L)).
All Firms
Number
Of Firms
(1)
93
AIP(1)
AIP(2)
AIP(3)
AIP(4)
61
30
22
23
0-5
(2)
2
2
0
1
1
Ownership Concentration (OWN(L))
5-10
10-20
20-40
40-50
(3)
(4)
(5)
(6)
23
28
19
7
20
12
5
6
20
8
8
5
12
8
2
7
3
2
1
2
.
50+
(7)
14
4
0
5
2
28
TABLE III
CORRELATION OF ANTI-INVESTOR PROTECTIONS
AND OWNERSHIP CONCENTRATION
The entries are the correlation coefficients between anti-investor protections (AIP's) and
measures of ownership concentration (OWN's). All of the AIP's are indicator variables taking a
value of 1 for the criteria below and zero elsewise: AIP(1)=1 if a firm issues preference shares;
AIP(2)=1 if the percentage of shares issued as tradable depository receipts is equal to or greater
than 50%; AIP(3)=1 if a firm issues priority shares; AIP(4)=1 if a firm is not required to
implement the structural regime, but does so voluntarily. Ownership concentration is measured
by the ownership stake of the largest shareholder stated as a percentage of all outstanding
ordinary equity (OWN(L)) or indicator variables, OWN(20) and OWN(40), taking a value of 1
for the criteria below and zero elsewise: OWN(20)=1 if OWN(L) > 0.20; OWN(40)=1 if
OWN(L) > 0.40. For a sample of 93 firms, the critical values for the null hypothesis of no
correlation are 0.170, 0.201, and 0.261 for the 10%, 5%, and 1% significance levels, respectively.
The figures in bold are those used to evaluate the substitution hypothesis.
AIP(1)
AIP(2)
AIP(3)
AIP(4)
OWN(L)
OWN(20)
OWN(40)
AIP(1)
(1)
1.000
0.112
-0.023
-0.005
-0.413
-0.348
-0.367
AIP(2)
(2)
AIP(3)
(3)
AIP(4)
(4)
1.000
-0.222
-0.022
-0.291
-0.167
-0.263
1.000
0.033
0.020
-0.063
0.062
1.000
-0.027
0.018
-0.071
OWN(L) OWN(20) OWN(40)
(5)
(6)
(7)
1.000
0.794
0.860
1.000
0.635
1.000
29
TABLE IV
PROFITABILITY REGRESSIONS
THE ROLE OF ANTI-INVESTOR PROTECTIONS
The parameter estimates are based on the following equation:
PROF = b0 + b1*AIP' + b2*CVCFA + b3*DIVERSITY
+ b4*LEVERAGE + b5*SIZE +
Σj dj*IDUMj + e,
where PROF is profitability (before-tax profits plus financial expenses) / TA, the latter defined as
total assets less depreciation; AIP' is the indicated combination in Panel A of the Anti-Investor
Protection variables -- AIP(1)=1 if a firm issues preference shares; AIP(2)=1 if the percentage of
shares issued as tradable depository receipts is equal to or greater than 50%; AIP(3)=1 if a firm
issues priority shares; CVCFA is the coefficient of variation (the standard deviation divided by
the mean) of CFA, the latter defined as cash flow (operating income plus depreciation) divided
by TA; DIVERSITY is the percentage of two-digit industrial activity in which the firm is
involved outside its core business, measured by revenues; LEVERAGE is (TA less stockholders
equity) / TA; SIZE is the natural logarithm of TA; IDUMj is a dummy for the jth industry; e is an
error term. The b's and d's are estimated by ordinary least squares. Standard errors in
-3
parentheses are heteroscedastic-consistent. RSS is the residual sum of squares raised to 10 .
(1)
A. Anti-Investor Protections
AIP(1) or AIP(2) or AIP(3)
(2)
(3)
(4)
2.408
(1.241)
AIP(1) or AIP(2)
0.282
(1.336)
AIP(1) or AIP(3)
2.184
(1.047)
AIP(2) or AIP(3)
B. Conditioning Variables
Constant
CVCFA
DIVERSITY
LEVERAGE
SIZE
C. Statistics
2
Adjusted R
RSS
(5)
2.776
(1.083)
18.409
(3.596)
-0.835
(2.104)
0.171
(0.283)
-10.609
(4.408)
0.337
(0.321)
16.598
(3.508)
-1.056
(2.203)
0.124
(0.264)
-9.229
(4.408)
0.243
(0.330)
18.142
(4.347)
-0.865
(2.171)
0.164
(0.291)
-10.339
(5.046)
0.325
(0.311)
17.367
(3.465)
-0.961
(2.163)
0.104
(0.266)
-9.285
(4.393)
0.247
(0.325)
14.678
(3.544)
-0.987
(2.011)
0.066
(0.270)
-6.058
(4.497)
0.216
(0.311)
0.166
1.540
0.197
1.465
0.157
1.539
0.196
1.466
0.237
1.392
30
TABLE V
PROFITABILITY REGRESSIONS
THE ROLE OF ANTI-INVESTOR PROTECTIONS
The parameter estimates are based on the following equation:
PROF = b0 + b1*AIP' + b2*CVCFA + b3*DIVERSITY
+ b4*LEVERAGE + b5*SIZE +
Σj dj*IDUMj + e,
where PROF is profitability (before-tax profits plus financial expenses) / TA, the latter defined as
total assets less depreciation; AIP' is the indicated combination in Panel A of the Anti-Investor
Protection variables -- AIP(1)=1 if a firm issues preference shares; AIP(2)=1 if the percentage of
shares issued as tradable depository receipts is equal to or greater than 50%; AIP(3)=1 if a firm
issues priority shares; AIP(4)=1 if a firm is not required to implement the structural regime, but
does so voluntarily; CVCFA is the coefficient of variation (the standard deviation divided by the
mean) of CFA, the latter defined as cash flow (operating income plus depreciation) divided by
TA; DIVERSITY is the percentage of two-digit industrial activity in which the firm is involved
outside its core business, measured by revenues; LEVERAGE is (TA less stockholders equity) /
TA; SIZE is the natural logarithm of TA; IDUMj is a dummy for the jth industry; e is an error
term. The b's and d's are estimated by ordinary least squares. Standard errors in parentheses are
-3
heteroscedastic-consistent. RSS is the residual sum of squares raised to 10 .
(1)
A. Anti-Investor Protections
AIP(4)
(2)
(3)
3.312
(1.318)
AIP(4) and AIP(1)
2.205
(1.527)
AIP(4) and AIP(2)
3.788
(2.113)
AIP(4) and AIP(3)
B. Conditioning Variables
Constant
CVCFA
DIVERSITY
LEVERAGE
SIZE
C. Statistics
2
Adjusted R
RSS
(4)
6.214
(2.861)
17.626
(3.309)
-0.325
(2.093)
0.068
(0.281)
-10.734
(4.115)
0.379
(0.304)
18.479
(3.465)
-0.625
(2.092)
0.162
(0.271)
-9.794
(4.466)
0.257
(0.304)
17.850
(3.567)
-0.177
(2.249)
0.205
(0.258)
-9.375
(4.529)
0.301
(0.301)
17.409
(2.992)
-1.028
(1.809)
0.084
(0.267)
-10.299
(4.018)
0.454
(0.274)
0.247
1.374
0.186
1.485
0.197
1.464
0.266
1.338
31
TABLE VI
PROFITABILITY REGRESSIONS
THE ROLE OF CONCENTRATED OWNERSHIP
The parameter estimates are based on the following equation:
PROF = b0 + b1*OWN' + b2*CVCFA + b3*DIVERSITY
+ b4*LEVERAGE + b5*SIZE +
Σj dj*IDUMj + e,
where PROF is profitability (before-tax profits plus financial expenses) / TA, the latter defined as
total assets less depreciation; OWN' is the indicated measure of concentrated ownership in Panel
A; CVCFA is the coefficient of variation (the standard deviation divided by the mean) of CFA,
the latter defined as cash flow (operating income plus depreciation) divided by TA; DIVERSITY
is the percentage of two-digit industrial activity in which the firm is involved outside its core
business, measured by revenues; LEVERAGE is (TA less stockholders equity) / TA; SIZE is the
natural logarithm of TA; IDUMj is a dummy for the jth industry; e is an error term. The b's and
d's are estimated by ordinary least squares. Standard errors in parentheses are heteroscedastic-3
consistent. RSS is the residual sum of squares raised to 10 .
A. Concentrated Ownership
OWN(L)
2
(1)
(2)
0.010
(0.023)
-0.041
(0.069)
0.001
(0.001)
OWN(L)
OWN(20)
(3)
-0.769
(1.027)
OWN(40)
B. Conditioning Variables
Constant
CVCFA
DIVERSITY
LEVERAGE
SIZE
C. Statistics
2
Adjusted R
RSS
(4)
0.315
(1.296)
18.030
(3.333)
-0.756
(2.167)
0.179
(0.295)
-10.826
(4.560)
0.364
(0.301)
19.067
(3.595)
-0.597
(2.198)
0.181
(0.292)
-10.615
(4.545)
0.306
(0.297)
19.159
(3.410)
-0.849
(2.065)
0.181
(0.276)
-10.280
(4.462)
0.282
(0.293)
18.230
(3.380)
-0.790
(2.160)
0.174
(0.291)
-10.676
(4.507)
0.353
(0.303)
0.158
1.536
0.152
1.527
0.161
1.530
0.157
1.538